> By moving to self-hosted, i was able to bring down my "cold-start"/"non-cached" wait time on tag filters from 30s to ~1s. I think highlighting the good perfo

> By moving to self-hosted, i was able to bring down my "cold-start"/"non-cached" wait time on tag filters from 30s to ~1s. I think highlighting the good performance is worth mentioning on your project's page (maybe with some stats). > Obviously, it's a different place in the design place. Hosting & deployment is much easier of course. Though workflow for adding new bookmarks seems more manual e.g. can you add new marks while on your mobile phone, quickly? This is exactly the reason why I did not write an easier way to add the bookmarks other than editing a text file. Because I don't want to end up having 67.000 bookmarks. I have an "inbox" folder in my browser where I put all the stuff I think is interesting. I later review them and either 1.) turn them into an actionable task in my todo manager 2.) delete them after finding out it's really not that interesting (most of them) or 3.) put it in my bookmarks file if I think this is something worth archiving. The last one does not happen very often, though. As you said, this is a much different use case than yours. I get how most people go your route, but it does not work for me. It encourages my fear of missing out: "oh, this is an article about XXX. let's archive it for later". But if I ever decide I want to learn about XXX, the article might be outdated and I might as well just search for a new up-to-date source (especially being a frontend developer, where yesterday's info is already outdated). However, I can see how such a system really shines when doing research.